Gnatcatcher

# Implementation status

This document outlines a new proposal to prevent covert tracking: Gnatcatcher.

• The Gnatcatcher proposal has entered public discussion.
• This proposal has not been implemented in any browser.
• The Privacy Sandbox timeline provides implementation timings for Gnatcatcher and other Privacy Sandbox proposals.

# Why do we need this proposal?

IP addresses were created to provide unique identifiers for a client so that traffic can be routed over the internet. IP addresses can be stable over periods of time, which can lead to user identification across first-parties.

Gnatcatcher, or the Global Network Address Translation Combined with Audited and Trusted CDN or HTTP-Proxy Eliminating Reidentification, is a proposed solution for IP blindness. This proposal suggests combining two other proposals: Willful IP Blindness and Near-Path NAT (or a proxy-related solution) for any connection not participating in Willful IP Blindness.

This means by default, IP addresses would be hidden. Sites may be allowed to attest they aren't misusing IP addresses to use direct connections.

# What is Willful IP Blindness?

The goal of Willful IP Blindness is to provide HTTP applications a mechanism to assert that they are not using IP addresses for cross-site tracking purposes.
The proposal also states this mechanism should allow for normal operations of servers, such as uses of IP addresses for bot, DoS, and SPAM detection.

Willful IP Blindness could be offered as a service by a content delivery network (CDN) or reverse proxies, to remove burden from the hosting providers.

# What is Near-Path Nat?

Near-Path NAT (Network Address Translator) proposes allowing groups of users to send traffic through the same server, thus leading all traffic to appear to originate from the same pool of IP addresses. This proposal suggests no server-side changes are necessary to guarantee IP address privacy, which is much easier to rollout for sites and users.

The browser will use Multiplexed Application Substrate over QUIC Encryption (MASQUE) to forward HTTP traffic through an IP privatizing server (IPPS). The HTTP traffic seen by servers will have the IP address of the IPPS instead of the IP address of the browser. To ensure the IPPS isn't privy to the HTTP traffic contents, the browser will use end-to-end encryption.

# How will Gnatcatcher work?

Gnatcatcher proposes use of both Near-Path NAT and Willful IP Blindness. Near-Path NAT (or another proxy solution) would use IP Addresses from destination servers as a baseline default.

There is a subset of web service providers who require additional controls to prevent abuse, who could choose to attest to Willful IP Blindness compliance so that the client can make a direct connection. That compliance may be ensured through audit and certification

# How will Willful IP Blindness be enforced?

There are a number of options in consideration for Willful IP Blindness enforcement. One possibility is to require an independent third-party coordinator for assessment and attestation to access IP addresses.

Enforcement for Gnatcatcher is still in discussion.

# When will Gnatcatcher be available?

The earliest date of scaled availability represents the earliest date when Gnatcatcher could be made available to sites for broad use on an optional basis. This will not happen before 2023.

At this time, Gnatcatcher is a proposal and has not been implemented for any browser.

# Engage and share feedback

The Gnatcatcher proposal is under active discussion and subject to change in the future. If you try this API and have feedback, we'd love to hear it.

• GitHub: Read the proposal, raise questions and participate in discussion.
• Developer support: Ask questions and join discussions on the Privacy Sandbox Developer Support repo.